scrumrun 1.5.2 → 2.0.0

package/templates/project/.scrumrun/runbook.md

@@ -1,101 +0,0 @@
-# ScrumRun Runbook
-
-## Purpose
-
-This runbook lets an AI agent execute one sprint independently and safely.
-
-## Natural-Language Intake
-
-When the owner describes work without a ScrumRun command and `Interaction Mode` is not `strict`:
-
-1. Identify the intent, expected outcome, urgency, scope, risk, and relationship to the main goal.
-2. Read only the canonical context and relevant history needed to verify the classification.
-3. Route the request to one of: quick task, knowledge/discovery, main sprint, corrective fix, backlog, isolated feature lane, or reject/defer.
-4. Present the recommendation, rationale, material unknowns, and at most two useful alternatives.
-5. Ask for approval when required by `.scrumrun/config.md`; otherwise state which configured policy authorizes the transition.
-6. Only after approval, invoke the matching workflow and preserve its normal safety rules.
-
-Intake may think and recommend autonomously. It must not interpret a vague acknowledgement as permission to execute application changes.
-
-## Required Start Sequence
-
-1. Read `AGENTS.md`.
-2. Read `.scrumrun/golden-rules.md`.
-3. Read `.scrumrun/config.md`.
-4. Read `.scrumrun/token-policy.md`.
-5. Read `.scrumrun/context.md` as a snapshot, not truth.
-6. Read `.scrumrun/map.md`.
-7. Read `.scrumrun/project.md` and `.scrumrun/knowledge.md`.
-8. Read `.scrumrun/runbook.md`.
-9. Read the relevant sprint plan:
-   - main goal: `.scrumrun/goals/main/sprint.md`
-   - feature lane: `.scrumrun/features/<feature>/sprint.md`
-10. Read the relevant history file.
-11. Check whether the requested sprint already has `Status: completed.`, `Status: partial.`, or `Status: blocked.`.
-12. If it already ran, ask whether to audit, rerun/fix, resume, or continue to another sprint.
-13. Identify dependencies, impacted files, env/config impact, database impact, and test plan.
-14. Only then implement.
-
-## Sprint Analysis Template
-
-```text
-Sprint:
-Lane: main goal | feature <slug>
-History status:
-Goal:
-Relevant docs/specs:
-Source areas to audit:
-External docs needed:
-Expected files:
-Env/config impact:
-Database impact:
-Test plan:
-Open questions:
-```
-
-## Execution Protocol
-
-1. Entenda
-2. Avalie Impactos
-3. Tire Duvidas
-4. Execute
-5. Teste
-
-## Safety Rules
-
-- Never modify read-only source paths.
-- Never commit real secrets.
-- Use env/config for runtime values.
-- Ask before changing behavior that is ambiguous or production-sensitive.
-- Use mocks/fakes for external providers in automated tests.
-- Keep main goal and feature lane histories separate.
-- Update `.scrumrun/context.md` when sprint execution changes current focus, risks, decisions, or map pointers.
-
-## History Update Template
-
-```md
-### YYYY-MM-DD - Sprint NN: Name
-
-Status: completed|partial|blocked
-
-Summary:
-- ...
-
-Files changed:
-- ...
-
-Commands run:
-- ...
-
-Tests/verification:
-- ...
-
-Env/config:
-- Added variable names only, no values: ...
-
-Decisions:
-- ...
-
-Risks/follow-ups:
-- ...
-```

package/templates/project/.scrumrun/token-policy.md

@@ -1,43 +0,0 @@
-# ScrumRun Token Policy - {{PROJECT_NAME}}
-
-This policy reduces token waste without weakening ScrumRun safety.
-
-## Rules
-
-1. Safety beats token economy. Never skip `core.md`, `golden-rules.md`, `config.md`, or relevant history to save tokens.
-2. `context.md` is a snapshot, not truth. Use it to choose what to read, then verify against canonical files.
-3. Prefer targeted reads over broad dumps. Read specific files, symbols, folders, and ranges when possible.
-4. Prefer summaries for large logs, generated output, dependency folders, old reviews, and old history.
-5. Do not read `.scrumrun/vault.local.md` unless the user explicitly asks for a local development credential workflow.
-6. Do not print vault values in summaries, context, history, decisions, reviews, or commits.
-7. Use `map.md` before repeated broad searches.
-8. Use approved knowledge before planning. Pending knowledge is not planning truth.
-9. When a source is skipped for token economy, say what was skipped and why.
-10. If uncertainty affects correctness or safety, read the source even if it costs more tokens.
-
-## Recommended Read Order
-
-```text
-AGENTS.md
-  -> core.md
-  -> golden-rules.md
-  -> config.md
-  -> token-policy.md
-  -> context.md
-  -> map.md
-  -> project.md
-  -> knowledge.md
-  -> runbook.md
-  -> current backlog / sprint / feature files
-  -> agents.md
-  -> relevant history + decisions
-  -> targeted source files
-```
-
-## Output Discipline
-
-- Summarize command output.
-- Include exact file references when they matter.
-- Avoid pasting large diffs unless requested.
-- For tests, report command, result, and relevant failures only.
-- For reviews, list findings by severity before summaries.